Projects

SegAgent-AI: A Dockerized AI Platform for Industrial Defect Segmentation, Severity Analysis, and Cloud Deployment

SegAgent-AI transforms a research-grade surface defect segmentation model into a deployable industrial inspection platform.

Code →

Forecasting Meteorological Time Series Using LSTM Networks

Developed machine learning applications for reference evapotranspiration and rainfall forecasting using weather-station data from Hawaii and Guam, with data pipelines and API integration for CropNet.

Demo / system →

Japanese Stock Market Forecasting Using LSTM and NLP

Built forecasting workflows using financial data and text-based market signals for approximately 300 Japanese companies to support trend discovery and investment-oriented analysis.

EEG Classification in BCI Applications

Designed dictionary-learning based EEG classification methods for brain-computer interface datasets, improving classification performance while avoiding expensive sparse optimization steps.

Paper →

ETT Localization in Chest X-Rays with U-Net++

Developed segmentation and localization models for endotracheal tube detection in chest radiographs using U-Net++ and compound loss functions, achieving high IoU on clinical data.

Code →

Fall Risk Assessment Using Hybrid ML

Designed machine-learning based classification systems for fall-risk assessment in older adults using computerized posturography and explainable AI methods.

Industrial Surface Defect Detection

Research and prototype work on CNN architectures for high-precision surface defect detection and segmentation in industrial inspection workflows.

Code →